Dev Genes Evol (2009) 219:289–300 DOI 10.1007/s00427-009-0290-z
ORIGINAL ARTICLE
Phylogenetic and evolutionary relationships and developmental expression patterns of the zebrafish twist gene family Gare Hoon Yeo & Felicia S. H. Cheah & Christoph Winkler & Ethylin Wang Jabs & Byrappa Venkatesh & Samuel S. Chong
Received: 9 January 2009 / Accepted: 30 April 2009 / Published online: 30 June 2009 # Springer-Verlag 2009
Abstract Four members of the twist gene family (twist1a, 1b, 2, and 3) are found in the zebrafish, and they are thought to have arisen through three rounds of gene duplication, two of which occurred prior to the tetrapod-fish split. Phylogenetic analysis groups most of the vertebrate Twist1 peptides into clade I, except for the Twist1b proteins of the acanthopterygian fish (medaka, pufferfish, stickleback), which clustered within clade III. Paralogies and orthologies among Communicated by M. Hammerschmidt Electronic supplementary material The online version of this article (doi:10.1007/s00427-009-0290-z) contains supplementary material, which is available to authorized users. G. H. Yeo : F. S. H. Cheah : B. Venkatesh : S. S. Chong (*) Department of Pediatrics, Yong Loo Lin School of Medicine, National University of Singapore, 5 Lower Kent Ridge Road, Singapore 119074, Singapore e-mail:
[email protected] C. Winkler Department of Biological Sciences, Faculty of Science, National University of Singapore, Singapore, Singapore
the zebrafish, medaka, and human twist genes were determined using comparative synteny analysis of the chromosomal regions flanking these genes. Comparative nucleotide substitution analyses also revealed a faster rate of nucleotide mutation/substitution in the acanthopterygian twist1b compared to the zebrafish twist1b, thus accounting for their anomalous phylogenetic clustering. We also observed minimal expression overlap among the four twist genes, suggesting that despite their significant peptide similarity, their regulatory controls have diverged considerably, with minimal functional redundancy between them. Keywords Twist . Zebrafish . Medaka . Phylogeny . Orthology . Synteny . Expression . Regulation Abbreviations hpf hours post-fertilization dpf days post-fertilization WISH whole-mount in situ hybridization UTR untranslated region
Introduction E. W. Jabs Departments of Genetics and Genomic Sciences, Pediatrics, and Developmental and Regenerative Biology, Mount Sinai School of Medicine, New York, USA B. Venkatesh Institute of Molecular and Cell Biology, Agency for Science, Technology, and Research, Biopolis, Singapore, Singapore S. S. Chong Children’s Medical Institute and Department of Laboratory Medicine, National University Hospital, Singapore, Singapore
Phylogenetic and gene expression studies are invaluable tools that aid in our understanding of the regulation and function of conserved genes and gene families. Together, they provide important clues to the evolutionary events and functional changes that have occurred in these genes in different species. The Twist gene is essential for development and survival, and is present in animals ranging from cnidarians and Drosophila to humans, either in single copy or as a gene family of two to five members. Twist was first isolated in Drosophila as a zygotic gene involved in the establishment of dorso-ventral patterning (Thisse et al. 1987). At gastrulation,
290
homozygous twist mutant embryos were abnormal and failed to differentiate their mesoderm (Simpson 1983; Thisse et al. 1987). Since this initial discovery, Twist orthologs and paralogs have been identified in many other animal species. Among them, Twist1 has been the most intensively studied, and its expression profile has been reported in many species including the mouse (Wolf et al. 1991; Fuchtbauer 1995; Stoetzel et al. 1995), rat (Bloch-Zupan et al. 2001), Xenopus (Hopwood et al. 1989; Stoetzel et al. 1998), chick (Tavares et al. 2001), medaka (Yasutake et al. 2004), and zebrafish (Rauch et al. 2003; Germanguz et al. 2007; Yeo et al. 2007). Twist2 (previously known as Dermo1) is another family member that is found in human (Lee et al. 2000), mouse (Li et al. 1995), rat (Maestro et al. 1999), chick (Scaal et al. 2001), medaka (Gitelman 2007), Fugu (Gitelman 2007), and zebrafish (NM_001005956). Its expression profile has been described in mouse (Li et al. 1995), chick (Scaal et al. 2001), and zebrafish (Thisse and Thisse 2004; Germanguz et al. 2007). In mouse, Twist2 is expressed in both the sclerotome and dermatome of the somite, the cranial mesenchymal cells around the nose, pharyngeal arches and tongue, whiskers, somites, limb, and branchial arches (Li et al. 1995). In chick, Twist2 is expressed in the somites, head mesenchyme, limbs, branchial arches, and mesenchyme of the feather buds (Scaal et al. 2001). A third family member Twist3 is absent in mammals but found in Xenopus, chick, medaka, stickleback, and zebrafish (Gitelman 2007). In contrast to Twist1 and Twist2, little is known about the role of Twist3. Adding to the complexity of this gene family, Twist2 or Twist3 are absent in some species while Twist1 or Twist3 paralogs are present in others. The different numbers of Twist genes in different species has been attributed to the gain or splitting of functional domains or regulatory sequences, or the merging of regulatory roles (Gitelman 2007). Germanguz et al. (2007) recently identified a fourth zebrafish twist gene (twist1a), and also described the expression patterns of all four genes. However, we observed significant differences between the reported expression profiles and our own expression studies. We now present a re-analysis of the expression profiles of all four members of the zebrafish twist gene family, and suggest a possible explanation for the observed differences with the published data.
Materials and methods Animal stocks and maintenance Singapore wild-type zebrafish were maintained in an aquarium system at 28°C. Embryos were obtained by
Dev Genes Evol (2009) 219:289–300
natural spawning and kept at 28°C and staged according to Kimmel et al. (1995) using standard morphological criteria. Older embryos were treated with 0.003% (2 nM) 1phenyl-2-thiourea to inhibit pigment formation and facilitate whole mount examination. Phylogenetic analysis A multiple sequence alignment of Twist protein sequences was first generated using Vector NTI® Suite 7.0. Gaps that were present in the alignment were removed before phylogenetic tree generation. A phylogenetic tree was generated by the neighbor-joining method (Saitou and Nei 1987) using PHYLIP software, with default gap opening penalty of 10 and gap extension penalty of 0.2, and with bootstrap analysis of 1,000 replicates. Phylogenetic analysis was performed with four zebrafish Twist peptides and 29 other Twist peptides from 11 species, based on GenBank cDNA sequences NM_000474 (human TWIST1), NM_057179 (human TWIST2), NM_053530 (rat Twist1), NM_021691 (rat Twist2), NM_011658 (mouse Twist1), NM_007855 (mouse Twist2), NM_204739 (chick Twist1), NM_204679 (chick Twist2), BK006265 (chick Twist3), NM_204084 (frog twist1), NM_001103209 (frog twist3), BK006281 (spotted green pufferfish twist1a), BK006282 (spotted green pufferfish twist1b), BK006283 (spotted green pufferfish twist2), NM_001104599 (fugu twist1a), NM_0011045 98 (fugu twist1b), NM_001104600 (fugu twist2), BK0062 68 (medaka twist1a), BK006269 (medaka twist1b), BK 006270 (medaka twist2), BK006271 (medaka twist3a), BK006272 (medaka twist3b), BK006276 (stickleback twist1a), BK006277 (stickleback twist1b), BK006278 (stickleback twist2), BK006279 (stickleback twist3a), BK006280 (stickleback twist3b), EF620930 (zebrafish twist1a), NM_130984 (zebrafish twist1b), NM_001005956 (zebrafish twist2), NM_130985 (zebrafish twist3), AF0979 14 (Japanese lancelet twist), and ABW34714 (sea urchin twist). Kimura 2-parameter nucleotide distances and Tajima–Nei-corrected number of substitutions per site at third codon positions were computed using MEGA 4.0 software (Tamura et al., 2007). Synteny analysis The map locations of the orthologous genes in human and zebrafish were obtained from NCBI Map Viewer (http:// www.ncbi.nlm.nih.gov/projects/mapview/), Human genome view (Build 36.3, March 2008) and Zebrafish genome view (Zv 7, July 2008), respectively. The map locations of Medaka twist1a (BK006268), twist1b (BK006269), twist2 (BK006270), twist3a (BK006271), and twist3b (BK0062 72) were determined using the search engine on the Ensembl Medaka Blast View (http://www.ensembl.org/
Dev Genes Evol (2009) 219:289–300
Oryzias_latipes/blastview). Other orthologous gene locations in Medaka were obtained from the Ensembl Oryzias latipes genome website (http://www.ensembl.org/Oryzias_ latipes/index.html; Ensembl release 50, July 2008). RT-PCR analysis Total RNA was extracted from embryos at various stages using TRIZOL® RNA isolation reagent (Invitrogen): onecell (0.2 hpf), eight-cell (1.25 hpf), 64-cell (2 hpf), 1,000cell (3 hpf), shield (6 hpf), bud (10 hpf), 14-somite (14 hpf), 1 day post-fertilization (dpf), 2 dpf, and 3 dpf. Intron-spanning gene-specific primer pairs for twist1a, twist1b, twist2, and twist3 transcript detection were 5′AGGTTCTACAGAGTGACGAGC-3′/5′-GCACAGGA TTCGAACTAGAGG-3′, 5′-TGTGGCGCACGAGAG ACT-3′/5′-GATCTATTCTGCATTGTGAC-3′, 5′-CGCA CGAGAGACTCAGTTAC-3′/5′-CCATACAGATAG CAGATAGCC-3′, and 5′-CTGAATCCCGAACTCT GATCC-3′/5′-GTGTTACCCGTCACTGAAG-3′, respectively. As a control for equivalent starting amounts of RNA template, β-actin transcript expression was detected using the primer pair 5′-CGCACTGGTTGTTGACAACG3′/5′-AGGATCTTCATGAGGTAGTC-3′. Total RNA was reverse-transcribed using SuperScript™ III reverse transcriptase (Invitrogen), followed by PCR using HotStarTaq™ DNA polymerase (Qiagen). Amplified products were analyzed by agarose gel electrophoresis. RNA in situ hybridization analysis Zebrafish twist1b and twist3 cDNA were amplified from reverse-transcribed RNA of 24 hpf embryos using the primer pairs 5′-CTGAACTCACTGGAAGGAGC-3′/5′GATCTATTCTGCATTGTGAC-3′ and 5′-CTGAATCCC GAACTCTGATCC-3′ /5′-C G AC ATC TC ATCC TAT TAGCG-3′, respectively. Amplified fragments were cloned into a pBlueScript®II KS(+) vector (Stratagene). Antisense RNA probes were synthesized from HindIII-linearized plasmid constructs using T7 RNA polymerase, while negative control sense RNA probes were synthesized from XbaI-linearized constructs using T3 RNA polymerase. Zebrafish twist1a and twist2 cDNA were amplified from reverse-transcribed RNA of 24 hpf embryos using the primers 5′-GAAAACACGAGGACCAATG-3′/5′-GAA TTGTACTAAAGCTTTGTA-3′ and 5′-GAACGGA CTGTTTACTTCCAC-3′/5′-CCATACAGATAGCAGAT AGCC-3′, respectively. Amplified fragments were cloned into a pCR®2.1-TOPO vector (Invitrogen). Antisense RNA probes were synthesized from EcoRV-linearized plasmid constructs using SP6 RNA polymerase, while negative control sense RNA probes were synthesized from BamHIlinearized constructs using T3 RNA polymerase.
291
Plasmids for two molecular markers, dlx2a and fli1a, were kindly provided by Monte Westerfield (University of Oregon) and Andrew D. Sharrocks (University of Manchester), respectively. Fluorescein-labeled RNA probes were prepared from linearized plasmids using the fluorescein RNA labeling kit (Roche Applied Sciences). Whole-mount in situ hybridization and tissue cryosection procedures were performed as described in Westerfield (2000). All images were captured using the MicroPublisher™ 5.0 color digital camera (QImaging™) and processed using Image-Pro® Plus version 5 software (Media Cybernetics).
Results Alignment, phylogenetic, and synteny analyses To understand the relationships of the zebrafish twist genes to each other and to those of a non-chordate, a cephalochordate, and other vertebrates, a phylogenetic tree was constructed using multiple alignments of homologous proteins belonging to human (Homo sapiens/Hs), rat (Rattus norvegicus/Rn), mouse (Mus musculus/Mm), chick (Gallus gallus/Gg), frog (Xenopus tropicalis/Xt), zebrafish (Danio rerio/Dr), medaka (Oryzias latipes/Ol), fugu (Takifugu rubripes/Tr), spotted green pufferfish (Tetraodon nigroviridis/ Tn), stickleback (Gasterosteus aculeatus/Ga), Japanese lancelet (Branchiostoma belcheri/Bb), and sea urchin (Lytechinus variegatus/Lv). The Twist proteins clustered into three clades, with the zebrafish Twist1 (Twist1a and Twist1b), Twist2 and Twist3 peptides falling into different clades (Fig. 1). Interestingly, zebrafish Twist1b did not cluster with the Twist1b of the other fish species medaka, fugu, pufferfish, and stickleback. We compared the genetic distance between twist1a and twist1b among these fish by using the Kimura 2-parameter model of nucleotide substitution, transitions, and transversions. The calculated Kimura value between zebrafish twist1a and twist1b was 0.26, whereas it was almost double in fugu, stickleback, pufferfish, and medaka (0.46, 0.47, 0.55, and 0.55 respectively). These values suggest a higher nucleotide substitution/mutation rate between the twist1a and twist1b sequences in the other fish species compared to the zebrafish. We also determined the Tajima–Nei-corrected number of substitutions per site at third codon positions between twist1a and twist1b sequences of zebrafish, fugu, pufferfish, stickleback, and medaka. Since most third-codon-position substitutions do not result in an amino acid change, mutations at these positions are assumed to be under little or no selective pressure and the calculated mutation rates are thus more likely to reflect neutral evolution. The Tajima–Nei value of 0.718 (±0.118) for zebrafish twist1a
292
Dev Genes Evol (2009) 219:289–300
Fig. 1 Rooted cladogram (a) and unrooted radial tree (b) of Twist proteins generated by the neighbor-joining method, with bootstrap values shown. The cladogram is rooted using sea urchin (Lv) as an outgroup. Zebrafish Twist1 (Dr1a and Dr1b), Twist2 (Dr2), and Twist3 (Dr3) proteins are clustered into the three major clades. Twist
protein members from human (Hs), rat (Rn), mouse (Mm), chick (Gg), frog (Xt), fugu (Tr), medaka (Ol), stickleback (Ga), pufferfish (Tn), lancelet (Bb), and sea urchin (Lv) were used to produce the phylogenetic trees
and twist1b was lower compared to fugu (1.113±0.248), stickleback (0.794±0.145), pufferfish (1.243±0.273), and medaka (0.992±0.183). These results again suggest a lower substitution/mutation rate between the twist1a and twist1b genes in the zebrafish compared with the other four fish species. Zebrafish twist1a (EF620930) is located on linkage group (LG) 16 at ~13.484 Mb, while twist1b (NM_130984) is located on LG19 at ~4.245 Mb, based on the Zv7 assembly, July 2007 release (http://www.ncbi.nlm.nih.gov/mapview/ map_search.cgi?taxid=7955). Both genes are linked to the hoxa gene cluster, whose duplication in ray-finned fishes was inferred to occur by whole genome duplication (Amores et al. 1998). In addition, several other genes are also duplicated on LG16 and LG19, with both chromosomal segments showing synteny to the human TWIST1 locus on chromosome 7p21 (~19.1 Mb; Supplementary Table 1). Thus, zebrafish twist1a and twist1b are co-paralogs and coorthologs of human TWIST1 that most likely arose through a fish-specific genome duplication event. Due to some confusion in the naming of the zebrafish LG16 and LG19 located twist1 co-paralogs in the genomic databases, we performed a comparative synteny analysis of the zebrafish and medaka twist1a and twist1b chromosomal
regions to resolve their orthologies. Medaka twist1a (BK006268) is located on LG16 at ~12.859 Mb, while twist1b (BK006269) is located on LG11 at ~8.931 Mb, based on the HdrR assembly, July 2008 release (http:// www.ensembl.org/Oryzias_latipes/index.html). When we compared the chromosomal region around zebrafish twist1a (on LG16) against the chromosomal regions around medaka twist1a (on LG16) and twist1b (on LG11), we observed a greater degree of conserved chromosomal synteny between zebrafish LG16 and medaka LG16 than between zebrafish LG16 and medaka LG11 (Supplementary Table 2). Conversely, when the chromosomal region around zebrafish twist1b (on LG19) was compared against the chromosomal regions around medaka twist1a and twist1b, we now observed a greater degree of conserved chromosomal synteny between zebrafish LG19 and medaka LG11 region than between zebrafish LG19 and medaka LG16 (Supplementary Table 3). These observations provide further confirmation that the twist1a and twist1b genes arose from a fish-specific genome duplication event and are true co-paralogs. Furthermore, the results indicate that the zebrafish LG16 and LG19 chromosomes are orthologous to the medaka LG16 and LG11 chromosomes, respectively.
Dev Genes Evol (2009) 219:289–300
293
To detect the presence of the four zebrafish twist transcripts at different embryonic stages, we first performed an RTPCR analysis of total RNA extracts using gene-specific intron-spanning amplification primers (Fig. 2a). Zebrafish twist1b and twist3 transcripts could be detected from as early as the one-cell stage, indicating the presence of
maternally expressed and deposited transcripts in the early embryo prior to zygotic transcript expression (Fig. 2b). In contrast, zebrafish twist1a and twist2 transcripts were detected only at or after the 1K-cell stage, suggesting that twist1a and twist2 transcripts are expressed exclusively from the zygote (Fig. 2b). These observations were further confirmed by wholemount in situ hybridization analysis (Fig. 3a–d). The four zebrafish twist genes share a high degree of nucleotide identity in their coding regions, especially between twist1a and twist1b, as well as in the highly conserved basic helix– loop–helix (bHLH) and tryptophan-arginine (WR) domains (Supplementary Table 6 and Supplementary Figure 1). Therefore RNA hybridization probes were synthesized to target unique 3′ untranslated regions (UTRs) of each transcript (Fig. 2a and Supplementary Figure 1). During gastrulation (5.25–10 hpf), twist1b and twist3 transcripts are ubiquitously expressed (Fig. 3f, h, j, l), whereas there was no detectable expression of twist1a (Fig. 3e, i). Interestingly, twist2 was specifically expressed in the organizer at the shield stage (6 hpf; Fig. 3g), and in the axial mesoderm at the bud stage (10 hpf; Fig. 3k). During early segmentation (2 to 5-somite stage, 10.7 to 11.7 hpf), twist1a transcripts were found to be concentrated in the premigratory neural crest cells (Fig. 3m, q). Zebrafish twist1b expression was detected in the head mesenchyme, sclerotome, and intermediate mesoderm (Fig. 3n, r; Yeo et al. 2007). Expression of twist2 was observed in the axial mesoderm (Fig. 3o, s). Zebrafish twist3 was still ubiquitously
Fig. 2 Reverse-transcription PCR detection of the zebrafish twist transcripts during embryonic development. a Structure of twist genes showing positions of intron-spanning primers (arrows) for RT-PCR and extent of unique 3′UTR probe targets (blue underline) for in situ hybridization. Empty boxes represent untranslated regions. Black boxes represent coding regions. GenBank accession numbers of
zebrafish twist1a, twist1b, twist2, and twist3 are NM_001017820, NM_130984, NM_001005956 and NM_130985, respectively. b RTPCR results showing twist1b and twist3 detection from the one-cell stage onwards, indicating the presence of maternally deposited transcripts. Zebrafish twist1a and twist2 are detectable only from the 1K-cell stage onwards, when zygotic transcription begins
Zebrafish twist2 (NM_001005956) is located on LG9 at ~39.145 Mb on the zebrafish Zv7 genome assembly. The zebrafish twist2 chromosomal segment is syntenic to the human and medaka twist2 locus on chromosome 2 and LG21 respectively (Supplementary Table 4). Zebrafish twist3 has no known mammalian ortholog, and lacks chromosomal synteny with any of the other twist loci in mammals, consistent with its placement in a clade separate from twist1 or twist2. It is located on LG23 at ~0.105 Mb on the zebrafish Zv7 genome assembly. In medaka, there are two twist3 genes, twist3a on LG5, and twist3b on LG7. Comparative synteny analysis revealed that the zebrafish LG23 region around twist3 shows a greater number of shared genes with medaka LG7 compared with medaka LG5 (Supplementary Table 5). The chromosomal synteny data indicate that medaka twist3a and twist3b are co-paralogs that arose through a chromosomal or genome duplication event, and that both are co-orthologs of zebrafish twist3. Embryonic expression patterns of the zebrafish twist gene family
294 Fig. 3 Whole-mount in situ RNA hybridization analysis of zebrafish twist gene expression during early embryogenesis. Lateral (a–p, u–x) and dorsal (q–t, y–ab) views are shown. Both twist1b and twist3 transcripts are detected ubiquitously from the cleavage (b, d) through gastrula (f, h, j, l) periods, while twist2 is first expressed in the organizer at the shield stage (g) and in the axial mesoderm at the bud stage (k). No twist1a transcripts were detectable from the cleavage (a) through gastrula (e, i) periods. During somitogenesis, twist1a is expressed in the premigratory neural crest cells (m, q, u, y), twist1b is expressed in the head mesenchyme, intermediate mesoderm and sclerotome (n, r, v, z), twist2 is expressed in the axial mesoderm, presumptive vasculature and caudal notochord (o, s, w, aa), and twist3 is expressed in the head mesenchyme and somites (t, x, ab). am axial mesoderm, cn caudal notochord, hm head mesenchyme, im intermediate mesoderm, nc neural crest cells, or organizer, pnc premigratory neural crest cells, pv presumptive vasculature, sc sclerotome, so somites
Dev Genes Evol (2009) 219:289–300
Dev Genes Evol (2009) 219:289–300
295
expressed at the 2-somite stage (Fig. 3p), but was expressed distinctly in the somites at the 5-somite stage (Fig. 3t). Later in the segmentation period (12-22 hpf), we observed both distinct and overlapping expression of the four zebrafish twist genes. At the 14-somite stage, twist1a transcripts were detected in the neural crest cells (Fig. 3u, y), and also in the sclerotome by late somitogenesis (Fig. 4i). Zebrafish twist1b was expressed in the head mesenchyme, intermediate mesoderm, and sclerotome (Fig. 3v, z). Localization of twist1b expression in the intermediate mesoderm was previously confirmed by two-color WISH using pax2.1 molecular marker (Yeo et al. 2007). Localization of twist1b expression in the head mesenchyme was confirmed by two-color in situ hybridization with dlx2a, a molecular marker for the adjacent forebrain (Fig. 5c). Zebrafish twist2 was expressed in the presumptive vascu-
lature and axial mesoderm (Fig. 3w, aa). With progressing somitogenesis, its expression in the axial mesoderm was greatly reduced, and was restricted to the caudal notochord by the 18-somite stage (18 hpf). Two-color in situ hybridization, using fli1a as a molecular marker for the presumptive vasculature, confirmed localization of twist2 in the presumptive vasculature (Fig. 5a, b, d). Similar to twist1b, expression of twist3 was observed in the head mesenchyme adjacent to the forebrain (Figs. 3x and 5e) as well as in the somites (Fig. 3ab). At the prim-5 stage (24 hpf), twist1a was expressed in the sclerotome, delaminated neural crest cells from the neural tube, and weakly in the fin bud (Fig. 4a, e, i). Zebrafish twist1b transcripts were localized in the head mesenchyme between the eye and diencephalon, cranial neural-crest-derived head mesenchyme, posterior somites,
Fig. 4 Expression of zebrafish twist genes during the pharyngula period. The prim-5 (a–i, k–m) and long-pec (j, n, o–v) stages are shown. Dorsal (a–d, o–r), lateral (e–h, s–v), transverse section (i, k, l, m), and flat mount (j, n) views are shown. At 24 hpf, twist1a is expressed in the sclerotome, neural crest cells delaminated from the neural tube and fin bud (a, e, i). Zebrafish twist1b is expressed in the neural-crest-derived head mesenchyme, tail bud and three preceding somites (b, f). No expression is observed along the trunk (k). twist2 expression is observed in the hypochord, dorsal aorta, and caudal notochord (c, g, l). twist3 is expressed in the head mesenchyme, fin
bud, and tail bud, with low expression in the somites (d, h, m). At 48 hpf, all four twist genes are expressed in the pharyngeal arches (o– v). Only twist1a and twist3 transcripts are detected in the fin buds (j, n, o, r, s, v). twist1a is also expressed in the fin fold and intermediate cell mass (s), and twist2 is weakly expressed in the dorsal aorta (q, u). Dotted lines denote the location of the transverse sections. at actinotrichs, cn caudal notochord, da dorsal aorta, ed endochondral disk, fb fin bud, ff fin fold, hc hypochord, hm head mesenchyme, icm intermediate cell mass, nc neural crest, no notochord, pa pharyngeal arches, sc sclerotome, tb tail bud
296
Dev Genes Evol (2009) 219:289–300
Fig. 5 Confirmation of zebrafish twist2 expression in the presumptive vasculature (a, b, d), and twist1a and twist3 expression in the head mesenchyme (c, e). Lateral (a, b), sagittal section (d, anterior to the left), and transverse section (c, e) views are shown. Except for a 14somite embryo in (a), the other results are from 18-somite embryos.
Double color in situ hybridizations with dlx2a molecular marker show that twist1b (c) and twist3 (e) are not expressed in the forebrain but in the adjacent head mesenchyme. twist2 is expressed in the presumptive vasculature (a, b, d), as shown by colocalization with molecular marker fli1a. fb forebrain, hm head mesenchyme, pv presumptive vasculature
and the tail bud (Fig. 4b, f, k; Yeo et al. 2007). Zebrafish twist2 transcripts were observed in the hypochord, dorsal aorta, and caudal notochord (Fig. 4c, g, l). Strong expression of twist3 was detected in the head mesenchyme, fin bud, and tail bud, with weak expression in the somites (Fig. 4d, h, m). By the long-pec stage (48 hpf), transcripts of all four zebrafish twist genes were detected in the pharyngeal arches (Fig. 4o–v). In addition, twist1a and twist3 transcripts were also observed in the pectoral fins, specifically in the actinotrichs for twist1a (Fig. 4j) and in the proximal (strong), and distal (weak) endochondral disk for twist3 (Fig. 4n). Expression of twist1a was also detected in the fin fold and intermediate cell mass (Fig. 4s). Weak expression of twist2 was detected in the dorsal aorta (Fig. 4q). At the protruding mouth stage (72 hpf), all four zebrafish twist transcripts continued to be observed in the pharyngeal arches (Fig. 6a-h). Additionally, twist1a was expressed in the head mesenchyme, heart valve, pectoral fin, fin fold and intermediate cell mass (Fig. 6a, e). Expression in the heart valve was not observed for the other twist genes (Fig. 6b–d). Zebrafish twist1b was also expressed in the olfactory placode (Fig. 6b, f; Yeo et al. 2007). Expression of twist3 was also detected in the pectoral fin (Fig. 6d, h).
Discussion Phylogenetic relationships of the twist genes in fish Using peptide sequences from 33 homologous Twist proteins belonging to 12 different species, a clear phylogenetic clustering of the vertebrate Twist proteins into three major clades was observed, each of which contained both teleost and tetrapod orthologs. Zebrafish Twist1a (Dr1a) and Twist1b (Dr1b) fall within clade I, together with Twist1a from medaka, stickleback, fugu, and pufferfish (Fig. 1). However, Twist1b from medaka, stickleback, fugu, and pufferfish did not cluster within clade I, but instead grouped together with the Twist3 proteins within clade III in the unrooted and rooted trees. These observations reflect the very high degree of peptide sequence similarity between Twist1a and Twist1b in zebrafish, a member of the superorder Ostariophysi, whereas the Twist1b peptides in the other four fish species, which all belong to the superorder Acanthopterygii, exhibit greater peptide similarity with Twist3 than with Twist1a (Supplementary Figure 2). While the Twist1b peptides of the acanthopterygii are tightly clustered together, their presence within clade III has raised questions regarding their true phylogenetic and evolutionary relationships with ostariophysian Twist1b. In
Dev Genes Evol (2009) 219:289–300
297
Fig. 6 Expression of zebrafish twist genes during the hatching period. Lateral (a–d) and dorsal (e–h) views are shown. At 72 hpf, all four twist genes are expressed in the pharyngeal arches (a–h). Zebrafish twist1a and twist3 are also expressed in the pectoral fins (e, h). Additionally, twist1a is expressed in the head mesenchyme, heart
valve (arrowhead), fin fold, and intermediate cell mass (a, e), while twist1b is also expressed in the olfactory placode (f). Insets show enlarged heart views. ff fin fold, hm head mesenchyme, hv heart valve, icm intermediate cell mass, of olfactory placode, pa pharyngeal arch, pf pectoral fin
fact, a recent phylogenetic analysis of a similar set of Twist peptides, using multiple phylogenetic reconstruction methods, placed the acanthopterygian Twist1b peptides outside of the three major clades (Gitelman 2007). The author postulated that this outgroup was likely an artifact of longbranch attraction, and felt that their placement within clade I was the most parsimonious choice. We now provide evidence from analysis of conserved synteny between the zebrafish and medaka twist1a and twist1b chromosomal regions that despite the observed anomalous phylogenetic grouping of the acanthopterygian Twist1b, these genes are the true orthologs of ostariophysian twist1b, as Gitelman (2007) had predicted. The comparative synteny studies also demonstrate that the ostariophysian and acanthopterygian twist1a and twist1b genes are co-paralogs and co-orthologs of mammalian Twist1/TWIST1. Thus, although phylogenetic analysis algorithms provide reasonable estimations of phylogenetic and evolutionary relationships of gene families among species, ambiguities may arise which could be resolved through other methods such as comparative analysis of conserved synteny. Our Kimura and Tajima–Nei analyses of the nucleotide substitution rates between twist1a and twist1b cDNA sequences are consistent with the observed closer evolutionary relationship among the four acanthopterygian fish as compared to the ostariophysian zebrafish. They also account for the observed phylogenetic clustering of the acanthopterygian Twist1b peptides into clade III. Given the substantial Twist1b peptide sequence difference between zebrafish and the other four fish species, twist1b expression and functional observations made in the zebrafish may not necessarily be extrapolatable to other species. Germanguz et al. (2007) have hypothesized that Twist1b in medaka,
stickleback, fugu, and pufferfish may be involved in building the interarcual cartilage, a cephalic neural crest derivative that defines all four fish. It would be interesting to see if the medaka twist1b expression profile has indeed diverged significantly from zebrafish. Our comparative synteny analyses also show that the sole zebrafish twist3 gene is orthologous to both medaka twist3a and twist3b, although there are more shared genes between the zebrafish twist3 chromosomal region and the medaka twist3b chromosomal region. The two medaka twist3 genes appear to have arisen through either a chromosomal or genome duplication event, as suggested by the presence of multiple shared genes between LG5 and LG7. Unlike twist1, however, the number of twist3 genes is variable, even within the acanthopterygii. There are two copies in medaka and stickleback, one in zebrafish, and none in fugu and pufferfish. Combining the data from our phylogenetic and comparative synteny analyses, we have reconstructed a model for the evolutionary history of the twist genes (Fig. 7). This model assumes the occurrence of three genome duplication events in the evolutionary history of the teleosts. The first two events occurred prior to the divergence of the teleost ancestors the ray-finned fishes (Actinopterygii) from the tetrapod ancestors the lobe-finned fishes (Sarcopterygii). In the first duplication event, the ancestral Twist gene probably gave rise to the Twist3 gene and a Twist1/Twist2 ancestral gene. This latter ancestral gene then split to give rise to Twist1 and Twist2 during the second duplication event. Subsequently during the teleost-specific third duplication event, which occurred approximately 350 million years ago (Ravi and Venkatesh 2008), the twist1 and twist3 genes were duplicated into twist1a and twist1b and into twist3a
298
Dev Genes Evol (2009) 219:289–300
Fig. 7 A model for the evolutionary history of the twist genes. Circles with a ‘D’ in the middle represent a gene or genome duplication event
and twist3b, respectively. The absence of a complete set of eight twist genes in any of the fish lineages that have been examined suggests that some of the duplicated genes were lost over time through functional redundancy. Common and unique expression sites of the zebrafish twist genes The mouse, chick, and Xenopus Twist genes are expressed in the pharyngeal arches (Hopwood et al. 1989; Wolf et al. 1991; Li et al. 1995; Scaal et al. 2001; Tavares et al. 2001). Similarly, we detected expression of all four zebrafish twist genes in the pharyngeal arches, corroborating the findings of others (Thisse and Thisse 2004; Germanguz et al. 2007). The Twist genes are also expressed in the somites or sclerotome in many species including mouse, chick, and Xenopus (Hopwood et al. 1989; Wolf et al. 1991; Li et al. 1995; Scaal et al. 2001; Tavares et al. 2001). Consistent with its expression in other species, we observed similar expression of zebrafish twist1a and twist1b in the sclerotome and twist3 in the somite. Interestingly, although twist2 appears to be expressed in the somite/sclerotome region, closer scrutiny indicates that it is in fact expressed in the presumptive vasculature during the segmentation period, and in the hypochord and dorsal aorta during the prim-5 stage (Figs. 3 and 4). Despite sharing some common expression sites, our studies reveal that the expression profiles of the four zebrafish twist genes otherwise differ significantly from each other. This includes even the highly homologous and most recently duplicated twist1a and twist1b genes. Zebrafish twist1a and twist1b are co-orthologs of tetrapod Twist1, which is present as a maternal transcript in early mouse and
Xenopus embryos (Stoetzel et al. 1995, 1998), and is expressed in the lateral plate mesoderm during early somitogenesis in the mouse, chick, and Xenopus (Hopwood et al. 1989; Wolf et al. 1991; Tavares et al. 2001). Using both RT-PCR and whole-mount in situ hybridization analyses, we first detected zebrafish twist1b expression in the zygote and early embryonic stages prior to the midblastula transition (1,000-cell stage), indicating that these are maternal transcripts (Figs. 2b and 3b). In contrast, we first detected zebrafish twist1a expression only at the 1,000cell stage by RT-PCR (Fig. 2b), with earliest tissue-specific expression in the premigratory neural crest cells at around the 2-somite stage (10.7 hpf) identified by WISH (Fig. 3m). The differences in zebrafish twist1a and twist1b expression profiles extend beyond the cleavage, blastula, and gastrula periods to the segmentation, pharyngula, and hatching periods. For example, during somitogenesis, twist1a is expressed in the neural crest cells (Fig. 3m, q, u, y) whereas twist1b is expressed in the head mesenchyme, sclerotome, and intermediate mesoderm (Fig. 3n, r, v, z). Other distinct examples include twist1a but not twist1b expression in the pectoral fin buds during the pharyngula and hatching periods, and twist1b but not twist1a expression in the tail bud at the prim-5 stage (Figs. 4 and 6). The zebrafish twist2 and twist3 genes also exhibit distinct expression profiles, with minimal similarities to each other or to twist1a or twist1b. For example, twist2 expression is first detected in the organizer during the shield stage, followed by the axial mesoderm in the bud stage (Fig. 3g, k). In contrast to twist1a and twist2 but similar to twist1b, twist3 is present as a maternal transcript (Fig. 2b). However, the similarity ends there, as zygotic twist3 expression is ubiquitous through the two-somite
Dev Genes Evol (2009) 219:289–300
299
generated RNA hybridization probes targeting the unique 3′ UTR sequences of each twist gene, which avoided the coding sequences entirely (Fig. 2a and Supplementary Figure 1). The earlier studies may have utilized RNA probes synthesized from full-length cDNAs or shorter ESTs (expressed sequence tags) or gene-specific fragments that contain coding sequence, which may explain the differences in detection sensitivity and specificity. Supplementary Table 7 summarizes the essential similarities and differences in expression patterns of the four zebrafish twist genes in this study and in other studies (Germanguz et al. 2007; Gitelman 2007). The observed common and unique expression sites of all four members of the zebrafish twist gene family indicate that twist gene members may perform both redundant and non-redundant functions during embryonic development. The functions of each gene member in a gene family may evolve over time through non-, neo-, sub-, and/or synfunctionalization (Postlethwait et al. 2004; Gitelman 2007; Roth et al. 2007). For example, the mouse twist1 expression domains in the nasal placode and cells of the limb bud are partitioned between zebrafish twist1b (olfactory placode) and zebrafish twist1a and twist3 (pectoral fin; Table 1). However, until the definitive expression patterns of these
stage (Fig. 3h, l, p). Other significant differences include the unique hypochord, dorsal aorta, and caudal notochord expression sites of twist2, and the pectoral fin bud and tail bud expression sites of twist3, all occurring during the pharyngula period (Fig. 4). The tail bud expression site of twist3 and twist1b appear to be similar. However, there appears to be a subtle difference in the pectoral fin bud expression sites of twist3 and twist1a, with twist3 expression concentrated in the endochondral disk (Fig. 4n) and twist1a expression strongest in the actinotrichs (Fig. 4j). While the expression patterns we observed in this study were generally similar to previous reports of one or more of the zebrafish twist genes (Rauch et al. 2003; Thisse and Thisse 2004; Germanguz et al. 2007), we also observed a number of significant differences in expression sites and developmental stages. These differences may be related to the use of different gene segments for RNA in situ hybridization probes in the different studies. Our alignment analysis of the four zebrafish twist cDNAs indicates a high degree of shared sequence similarity in their coding regions, especially between twist1a and twist1b, as well as within the bHLH and WR domains of all four genes (Supplementary Table 6). As in our previous analysis of zebrafish twist1 expression (Yeo et al. 2007), we therefore
Table 1 Twist expression sites in selected species Mm1a
Mm2b
Gg1c
Gg2d
Lateral plate mesoderm Head mesenchyme (neural-crest-derived) Intermediate mesoderm Somite Caudal gut Neural crest cells
✓ ✓
✓
✓ ✓
✓
✓
✓
Olfactory placode Pharyngeal arch/branchial arch Axial mesoderm/caudal notochord Tail bud Pectoral fin bud/limb buds Hypochord Dorsal aorta Fin fold /body wall Presumptive vasculature Heart valve Intermediate cell mass
✓ ✓
✓
✓
✓
✓ ✓
✓
✓
✓
a
Wolf et al. 1991
b
Li et al. 1995
c
Tavares et al., 2001
d
Scaal et al. 2001
e
Hopwood et al. 1989
f
Yeo et al. 2007
Xl1e
Dr1a
Dr1bf
✓
✓ ✓ ✓ ✓
Dr2
✓ ✓ ✓ ✓
✓
✓
✓
✓
✓
✓ ✓ ✓
✓
✓ ✓
✓ ✓
✓ ✓
✓ ✓ ✓ ✓
✓
Dr3
✓ ✓ ✓ ✓
✓ ✓ ✓ ✓
300
genes are confirmed, it would not be easy to make meaningful conclusions about the evolutionary changes that have occurred in the various zebrafish twist genes. Additional detailed and careful expression profiles of all the Twist genes in other species are necessary in order to construct a comprehensive understanding of the evolution of the vertebrate, and teleost, twist genes. Nevertheless, what is clear is that despite the still significant degree of peptide similarity among the four zebrafish twist genes, their regulatory control has diverged to the point that there is now minimal overlap of their developmental expression profiles, and thus minimal functional redundancy among them. Acknowledgments We thank Vladimir Korzh, Jin Ben (Institute of Molecular and Cell Biology, Singapore), and Karuna Sampath (Temasek Life Science Laboratories, Singapore) for invaluable advice, Monte Westerfield and Andrew D. Sharrocks for their gifts of dlx2a and fli1a (pAS160) plasmids respectively. This work was supported by grants R-178-000-104-112 (to SSC) from the National University of Singapore Academic Research Fund, and BMRC 06/1/21/19/459 (to SSC) and BMRC 07/1/21/19/544 (to CW) from the Biomedical Research Council of the Agency for Science, Technology, and Research, Singapore.
References Amores A, Force A, Yan YL, Joly L, Amemiya C, Fritz A, Ho RK, Langeland J, Prince V, Wang YL, Westerfield M, Ekker M, Postlethwait JH (1998) Zebrafish hox clusters and vertebrate genome evolution. Science 282:1711–1714 Bloch-Zupan A, Hunter N, Manthey A, Gibbins J (2001) R-twist gene expression during rat palatogenesis. Int J Dev Biol 45:397–404 Fuchtbauer EM (1995) Expression of M-twist during postimplantation development of the mouse. Dev Dyn 204:316–322 Germanguz I, Lev D, Waisman T, Kim CH, Gitelman I (2007) Four twist genes in zebrafish, four expression patterns. Dev Dyn 236: 2615–2626 Gitelman I (2007) Evolution of the vertebrate twist family and synfunctionalization: a mechanism for differential gene loss through merging of expression domains. Mol Biol Evol 24: 1912–1925 Hopwood ND, Pluck A, Gurdon JB (1989) A Xenopus mRNA related to Drosophila twist is expressed in response to induction in the mesoderm and the neural crest. Cell 59:893–903 Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF (1995) Stages of embryonic development of the zebrafish. Dev Dyn 203:253–310 Lee MS, Lowe G, Flanagan S, Kuchler K, Glackin CA (2000) Human Dermo-1 has attributes similar to twist in early bone development. Bone 27:591–602
Dev Genes Evol (2009) 219:289–300 Li L, Cserjesi P, Olson EN (1995) Dermo-1: a novel twist-related bHLH protein expressed in the developing dermis. Dev Biol 172:280–292 Maestro R, Dei Tos AP, Hamamori Y, Krasnokutsky S, Sartorelli V, Kedes L, Doglioni C, Beach DH, Hannon GJ (1999) Twist is a potential oncogene that inhibits apoptosis. Genes Dev 13:2207–2217 Postlethwait J, Amores A, Cresko W, Singer A, Yan YL (2004) Subfunction partitioning, the teleost radiation and the annotation of the human genome. Trends Genet 20:481–490 Rauch GJ, Lyons DA, Middendorf I, Friedlander B, Arana N, Reyes T, Talbot WS (2003) Submission and curation of gene expression data. ZFIN direct data submission Ravi V, Venkatesh B (2008) Rapidly evolving fish genomes and teleost diversity. Curr Opin Genet Dev 18:544–550 Roth C, Rastogi S, Arvestad L, Dittmar K, Light S, Ekman D, Liberles DA (2007) Evolution after gene duplication: models, mechanisms, sequences, systems, and organisms. J Exp Zoolog B Mol Dev Evol 308:58–73 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425 Scaal M, Fuchtbauer EM, Brand-Saberi B (2001) cDermo-1 expression indicates a role in avian skin development. Anat Embryol (Berl) 203:1–7 Simpson P (1983) maternal–zygotic gene interactions during formation of the dorsoventral pattern in Drosophila embryos. Genetics 105:615–632 Stoetzel C, Weber B, Bourgeois P, Bolcato-Bellemin AL, Perrin-Schmitt F (1995) Dorso-ventral and rostro-caudal sequential expression of Mtwist in the postimplantation murine embryo. Mech Dev 51:251–263 Stoetzel C, Bolcato-Bellemin AL, Bourgeois P, Perrin-Schmitt F, Meyer D, Wolff M, Remy P (1998) X-twi is expressed prior to gastrulation in presumptive neurectodermal and mesodermal cells in dorsalized and ventralized Xenopus laevis embryos. Int J Dev Biol 42:747–756 Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 24:1596–1599 Tavares AT, Izpisuja-Belmonte JC, Rodriguez-Leon J (2001) Developmental expression of chick twist and its regulation during limb patterning. Int J Dev Biol 45:707–713 Thisse B, Thisse C (2004) Submission and curation of gene expression data. ZFIN direct data submission. Thisse B, el Messal M, Perrin-Schmitt F (1987) The twist gene: isolation of a Drosophila zygotic gene necessary for the establishment of dorsoventral pattern. Nucleic Acids Res 15:3439–3453 Westerfield M (2000) The Zebrafish book: a guide for the laboratory use of zebrafish (Danio rerio). Wolf C, Thisse C, Stoetzel C, Thisse B, Gerlinger P, Perrin-Schmitt F (1991) The M-twist gene of Mus is expressed in subsets of mesodermal cells and is closely related to the Xenopus X-twi and the Drosophila twist genes. Dev Biol 143:363–373 Yasutake J, Inohaya K, Kudo A (2004) Twist functions in vertebral column formation in medaka, Oryzias latipes. Mech Dev 121:883–894 Yeo GH, Cheah FS, Jabs EW, Chong SS (2007) Zebrafish twist1 is expressed in craniofacial, vertebral, and renal precursors. Dev Genes Evol 217:783–789
